Here's US employment by sector.
This isn't a criticism of you, as you obviously linked for the purpose of the "real" (I'm sure the "real" data is also massaged, but less so than the forecasts) data in your link, but I love how full of shit those projections are. +3.1m jobs 2002-2012, +15.6m 2012-2022? Hahahaha. Goods-producing, excluding agriculture: -4.1m/+1.2m (somehow it will take more humans to produce goods than before!); Services-providing: +7.5m/+14.1m (of course services won't be made more efficient and automated; we will need far more humans than before!), Nonagriculture self-employed and unpaid family worker: -0.2/+0.5m (perhaps they have a non-fantasy/non-biased reason for this one, as it's nearly meaningless to the overall numbers). Only agriculture, forestry, fishing, and hunting admits that trends are likely to accelerate, giving -0.1m/-0.2m. I assume this is because nobody has those jobs anyway, so they couldn't pretend to forecast a large enough gain to really affect their overall numbers.